Near-surface Structure Imaging With Distributed Acoustic Sensing Technology

For a long time,earthquakes,landslides,and ground subsidence have been affecting people’s life and property safety.High-resolution shallow subsurface structures can be used for reliable seismic simulation and detection of underground mined-out areas beneath cities,which is of great significance for earthquake prevention and disaster reduction.Seismic tomography is an important means of constructing fine surface structures.In recent years,significant progress has been made in high-resolution imaging of urban and fault zone structures based on dense arrays composed of shortperiod seismometers.However,as the observation density further increases,there are significant challenges in conducting dense observations based on traditional seismometers.With the development of fiber optic sensing technology,a new observation technology,distributed acoustic sensing(DAS),has begun to be used in seismology studies.DAS directly uses fiber optic cables as sensors and can achieve ultra-high-density long-term observation at low cost,and is widely used in seismological research in multiple scenarios.This article focuses on the imaging of near-surface structures using seismic signals collected by DAS.(1)High-frequency surface wave imaging is one of the commonly used methods for near-surface structure imaging.In order to discuss the reliability of DAS for highfrequency surface wave imaging,we laid a 600-meter optical cable in a controllable environment,recorded the surface wave signals excited by an active source hammer and extracted clear dispersion curves for two-dimensional S-wave velocity structure inversion.The comparison with the results of co-located geophone recordings and ambient noise tomography shows the reliability of the results.(2)One of the advantages of DAS is that it can be combined with existing communication cables for observation,which can further reduce observation costs.We used an existing 5.2-kilometer-long city communication cable to conduct observations of large-volume airgun signals on the outskirts of the city and discussed the reliability of recording signals with existing optical cables.Due to the influence of strong noise levels and weak coupling,we proposed increasing the gauge length,multi-fiber stacking,and using engineered optical cables to improve the signal-to-noise ratio recorded by DAS,providing ideas for shallow structure imaging based on urban optical cables.(3)In urban environments,conducting active source imaging can affect residents’daily lives and is subject to strong noise interference.Therefore,we carried out ambient noise tomography on the urban fiber-optic cable recordings.After 4-hour stacking,we obtained the stable noise cross-correlation functions.The results showed that we constructed a two-dimensional S-wave velocity profile for some sections and revealed the presence of a low-velocity anomaly,while for other sections,reliable dispersion curves were difficult to extract.Through synthetic tests,we found that this may be due to the uneven distribution of noise sources.We estimated the distribution of noise intensity along the cable,which good correspondence to the intersection locations.(4)After clarifying the impact of uneven noise source distribution on ambient noise tomography results,we employed three station interferometric techniques to suppress the precursor wave signals in the noise cross-correlation function and improve the signal-to-noise ratio of surface waves.For the sections that were difficult to image,we successfully imaged the S-wave velocity structure below the road and revealed the location of hidden faults beneath the urban area.We separately discussed the effect of each of the three station interferometric techniques on improving the time resolution of ambient noise tomography and the accuracy of extracting dispersion curves.We also demonstrated that this method remains effective in horizontally non-uniform media.(5)The imaging results presented above are all based on dispersion curve inversion.Since DAS records axial single-component strain signals,it cannot separate different seismic phases like traditional seismometers through component rotation.Therefore,in the processing,we generally keep the source and receiver arrangements in the same line,where DAS records only Rayleigh wave signals,and we extract the dispersion curve of Rayleigh waves.When the source and receiver arrangements are not in the same line,DAS records a mixed wavefield of Rayleigh and Love waves,which affects the extraction of dispersion curves and the utilization rate of data.We propose to abandon the extraction of dispersion curves and directly use the dispersion spectrum of the mixed wavefield for inversion,and verify the reliability of the spectral inversion method through synthetic tests.For three sets of data collected from the same cable section,we obtained consistent results using the spectrum inversion method,demonstrating its feasibility in real data.Finally,we discuss the impact of various factors on the spectrum inversion method separately and demonstrate the stability of the method.Overall,this study explores surface wave imaging using DAS with different seismic sources and fiber-optic cable configurations.The high spatial resolution of DAS is effectively leveraged to achieve precise imaging of shallow structures beneath urban areas,revealing small-scale structural anomalies.We also tackles some of the limitations and challenges of using DAS in surface wave imaging,including low signalto-noise ratio,non-uniform noise source distribution affecting,and the inability to separate wavefields using single-component strain data.Possible solutions to these challenges are proposed,providing valuable insights for future research in this field.